ABSTRACTInflation as an economic indicator is very interesting to be discussed, although the author of the predecessor has done extensive research on the factors affecting inflation in Indonesia, the authors are interested to use ECM (Error correction model) in the relationship between inflation and the central bank interest rate, money supply, and the exchange rate. Time series data related to finance generally fluctuate and are not stationary at degree level, but the data from December 1997 until December 2011 found that only one variable that is not stationary at the level which is the central bank interest rate, but the first difference of all variables that studied already stationary. Once the data is tested cointegration with Granger 2-Step Engle concluded that the cointegrated variables can be made. First ECM results unsatisfactory estimation, this is due to the persistence of autocorrelation in ECM. The treatment of autocorrelation with GLS provide a better estimate, with a sign of coefficients according to economic theory, although variable m2 or the amount of money in circulation do not significantly affect the short-term relationship.Keywords: Inflation, The Central Bank Interest Rate, Money supply, Exchange rate, Error Correction Model